grandrea / protpipe

README

This repository facilitates quick, simple, and reproducible access to Data Independent Acquisition (DIA) proteomics workflows with minimal command line experience.

We include a convenient wrapper script for running DIA-NN inside a pre-built singularity image to first estimate protein abundance from raw mass spec output. Protein abundance estimates (accepting estimates from both DIA-NN and Spectronaut) can be processed in a preconfigured R environment, generating QC reports, and various analyses and visualizations.

Quick-start

1. Ensure singularity is installed and accessible on your system. Many HPCs (including NIH Biowulf) come with this pre-installed as a module. If your HPC has singularity installed, it will be automatically detected and loaded when necessary.
2. Clone this repository, i.e. execute git clone https://github.com/lzy604/Protpipe.git
3. If you are predicting protein abundances from raw mass spec output, look over and edit any custom DIA-NN parameters inside config.txt. You can either edit config.txt directly (and it will be used by default), or make a copy and save it to a different file name, then reference it with --cfg newfilename.txt when running the wrapper script.

Installing Singularity

This workflow requires that Singularity be available, which runs natively on a Linux system. Singularity is containerization software that allows an entire pre-configured computing environment to be accessed--reducing installation headaches and improving reproducibility.

We highly recommend using a workstation or HPC with a native Linux installation. Not only does this simplify the usage of singularity, it also would likely provide greater resources for DIA-NN's intensive computation.

To run on your personal/local non-Linux machine, Mac users need to first install a number of dependencies. Windows users would either need to use a virtual machine, or run things through the Windows Subsystem for Linux (WSL). Explaining the installation of singularity on these non-Linux systems is beyond the scope of this guide, so we defer to the documentation here.

Predicting Protein Abundances (running DIA-NN)

After editing the contents of config.txt, or generating a new file to specify with --cfg newfile.txt:

# Submit to SLURM
sbatch src/diann.sh --cfg config.txt

# Run Locally
src/diann.sh --cfg config.txt

Post-analysis

Running containerized interactive R session

You can start an interactive R session within the container as follows:

./protpipe.sh R

The above is shorthand for executing the followiing:

singularity exec -B ${PWD} R

where singularity exec R starts the R session, while -B ${PWD} binds the current directory within the container. Without binding, the current directory's files would not be visible inside the container.

Basic MS data analysis: ./protpipe.sh basic

For performing QC and running differential abundance or enrichment analysis for typical mass spec data. The required inputs are

• protein intensity estimates from DIA-NN or Spectronaut
• experimental design matrix csv file

./protpipe.sh basic \
    --pgfile EXAMPLES/DIFF_ABUNDANCE/iPSC.csv \
    --design EXAMPLES/DIFF_ABUNDANCE/design_matrix_iPSC.csv \
    --out EXAMPLES/DIFF_ABUNDANCE/

Affinity Purification Mass Spec analysis: ./protpipe.sh APMS

Similar to basic but for affinity purification mass spec. Requires the user to specify which protein was used for pulldown (--ip)

./protpipe.sh APMS \
    --pgfile EXAMPLES/APMS/APMS.csv \
    --design EXAMPLES/APMS/design_matrix_APMS.csv \
    --ip UNC13A \
    --out EXAMPLES/APMS/

Immunopeptidome analysis: ./protpipe.sh immuno

Requires the csv or tsv output from FragPipe and a csv specifying HLA typing.

./protpipe.sh immuno \
    --pepfile EXAMPLES/IMMUNOPEPTIDOME/combined_peptide.tsv \
    --out EXAMPLES/IMMUNOPEPTIDOME/ \
    --hla EXAMPLES/IMMUNOPEPTIDOME/HLA_typing.csv

Converting Mass Spec file formats

DIA-NN cannot handle some propietary file formats such as thermo fisher RAW. Thus these files must be converted (i.e. to mzML) prior to running DIA-NN. Conversion can be done interactively or by submitting to your HPC with sbatch.

Converting a single file:

outdir='mzML'
mkdir ${outdir}
bash src/pwiz-convert.sh /path/to/your/MassSpecFile.raw ${outdir}

Mass spec file conversion is handled by ProteoWizard (via wine in a singularity container). A writable sandboxed version of the container (which is required to run ProteoWizard) was built and modified from a docker image on March 02 2023. Steps were modified from here.

Building pwiz container

# Build writable singularity sandbox image based on docker image
singularity build --sandbox pwiz_sandbox docker://chambm/pwiz-skyline-i-agree-to-the-vendor-licenses

# Modified pwiz_sandbox/usr/bin/mywine
echo """#!/bin/sh

GLOBALWINEPREFIX=/wineprefix64
MYWINEPREFIX=/mywineprefix/

if [ ! -L "$MYWINEPREFIX"/dosdevices/z: ] ; then 
  mkdir -p "$MYWINEPREFIX"/dosdevices
  cp "$GLOBALWINEPREFIX"/*.reg "$MYWINEPREFIX"
  ln -sf "$GLOBALWINEPREFIX/drive_c" "$MYWINEPREFIX/dosdevices/c:"
  ln -sf "/" "$MYWINEPREFIX/dosdevices/z:"
  echo disable > $MYWINEPREFIX/.update-timestamp        # Line being added
  echo disable > $GLOBALWINEPREFIX/.update-timestamp    # Line being added
fi 

export WINEPREFIX=$MYWINEPREFIX
wine "$@"
""" > pwiz_sandbox/usr/bin/mywine

tar -czvf pwiz_sandbox.tar.gz pwiz_sandbox
rclone copy pwiz_sandbox.tar.gz onedrive:/singularity       # upload archive to cloud

Bulk convert raw to mzml

First, create rawfiles.txt that contains all .raw files to convert (one per line)

For example:

# Define file name
filelist='rawfiles.txt'
outdir='mzML'
mkdir ${outdir}

# find all files within the folder `ANXA11_redux` and print to file
cat <(find ANXA11_redux | grep raw) > ${filelist}

# get number of files
nfiles=$(wc -l ${filelist} | awk '{print $1}')

# Submit array, one per file
sbatch --array=1-${nfiles} src/pwiz-convert-array.sh ${filelist} ${outdir}